Sunday, August 17, 2008

Cardiovascular Risk Factors on Kitava, Part III: Insulin

The Kitava study continues to get more and more interesting in later publications. Dr. Lindeberg and his colleagues continued exploring disease markers in the Kitavans, perhaps because their blood lipid findings were not consistent with what one would expect to find in a modern Western population with a low prevalence of CVD.

In their next study, the researchers examined Kitavans' insulin levels compared to Swedish controls. This paper is short but very sweet. Young Kitavan men and women have a fasting serum insulin level considerably lower than their Swedish counterparts (KM 3.9 IU/mL; SM 5.7; KW 3.5; SW 6.2). Kitavan insulin is relatively stable with age, whereas Swedish insulin increases. In the 60-74 year old group, Kitavans have approximately half the fasting serum insulin of Swedes. One thing to keep in mind is that these are average numbers. There is some overlap between the Kitavan and Swedish numbers, with a few Kitavans above the Swedish mean.

In figure 2, they address the possibility that exercise is the reason for Kitavans' low insulin levels. Kitavans have an activity level comparable to a moderately active Swedish person. They divided the Swedes into three categories: low, medium, and high amounts of physical activity at work. The people in the "low" category had the highest insulin, followed by the "high" group and then the "medium" group. The differences were small, however, and Kitavans had far lower serum insulin, on average, than any of the three Swedish groups. These data show that exercise can not explain Kitavans' low insulin levels.

The researchers also found that they could accurately predict average Swedish and Kitavan insulin levels using an equation that factored in age, BMI and waist circumference. This shows that there is a strong correlation between body composition and insulin levels, which applies across cultures.

Now it's time to take a step back and do some interpreting. First of all, this paper is consistent with the idea (but does not prove) that elevated insulin is a central element of overweight, vascular disease and possibly the other diseases of civilization. While we saw previously that mainstream blood lipid markers do not correlate well with CVD or stroke on Kitava, insulin has withstood the cross-cultural test.

In my opinion, the most important finding in this paper is that a high-carbohydrate diet does not necessarily lead to elevated fasting insulin. This is why I think the statement "carbohydrate drives insulin drives fat" is an oversimplification.  With a properly-functioning pancreas and insulin-sensitive tissues (which many people in industrial societies do not have), a healthy person can eat a high-carbohydrate meal and keep blood glucose under control. Insulin definitely spikes, but it's temporary. The rest of the day, insulin is at basal levels. The Kitavans show that insulin spikes per se do not cause hyperinsulinemia.

So this leads to the Big Question: what causes hyperinsulinemia?? The best I can give you is informed speculation. Who has hyperinsulinemia? Industrial populations, especially the U.S. and native populations that have adopted Western foods. Who doesn't? Non-industrial populations that have not been affected by Western food habits, including the traditional Inuit, the Kuna, the traditional Masai and the Kitavans.

We can guess that total fat, saturated fat and carbohydrate do not cause hyperinsulinemia, based on data from the Inuit, the Masai and the Kitavans, respectively. We can also guess that there's not some specific food that protects these populations, since they eat completely different things. Exercise also can not completely account for these findings. What does that leave us with? Western food habits. In my opinion, the trail of metabolic destruction that has followed Westerners throughout the world is probably due in large part to industrial foods, including refined wheat flour, sugar and seed oils.

I'm not the first person to come up with this idea, far from it. The idea that specific types of carbohydrate foods, rather than carbohydrate in general, are responsible for the diseases of civilization, has been around for at least a century. It was an inescapable conclusion in the time of Weston Price, when anthropologists and field physicians could observe the transitions of native people to Western diets all over the world. This information has gradually faded from our collective consciousness as native cultures have become increasingly rare. The Kitava study is a helpful modern-day reminder.


Julie G said...

Super interesting!
So what you're trying to say overall is I can eat all the butter and cream I want as long as it's grass fed and preferably raw... and as long as someone hasn't snuck in some wheat?
Thanks Fat Boy!

Pastor Christopher Sylvain said...

The removal of the bran layer causes faster release leading to a higher insulin spike. The resultant liponeogenesis occurs in an exagerated fashion. Whole grains, fruits, vegetables all contain fiberous substances that slow glucose release leading to a much lower and slower insulin release. God provided the foods we need without needing major alterations.

Jeff said...


Great post, as usual.

Low-carb does look like a simplification. The details about which carbs appears more important. Low-carb, since it avoids the problem stuff, will get you around the problem. Eating the "right" carbs would also work.

Can you provide more detail about the types of tubers they eat? Is it like our potato, or is the potato too starchy and not fibrous enough?



Stan Bleszynski said...

I have a slightly diffrent speculation on this subject, see my Aug-15 post:

Stephan Guyenet said...

Hi Julie,

That's pretty much what I'm saying! Tell your cardiologist that some blog told you to eat unlimited grass-fed butter and hold the wheat. Make sure to send me a photo of his expression.

Stephan Guyenet said...

Hi Jeff,

That's pretty much exactly what I'm thinking. Low-carb sidesteps the problem of a damaged metabolism, while reducing the consumption of the biggest offenders.

Kitavans eat sweet potatoes, cassava, taro and yams. I don't see any reason to think that their root crops are qualitatively different from ours (sweet potato, potato). Yams, sweet potatoes, taro and cassava are all very starchy. Taro actually has more starch per weight than a potato. The glycemic index/load of these are not particularly low.

By the way, I checked out your blog, looks interesting! Sounds like you've made some good progress. Glad to have you on the blog.

Stephan Guyenet said...

Hi Stan,

If I understand correctly, you're saying that snacking is the problem? One thing that I've noticed is that French people don't snack. My Mom is convinced that's a big part of why they're healthy. Maybe so.

The problem is it's hard to disentangle correlation from cause here. Are people overweight because they snack, or do they snack because their metabolism is screwy?

I think it's worth pointing out that the Kitavans' root vegetables are not particularly low GI/GL. They eat a lot of them, and don't eat much fat to slow down digestion. Makes me think we're able to handle it if everything is running smoothly.

Anna said...

I can assure Christopher that fiber in whole grain, fruit and vegetables doesn't slow glucose release enough to prevent roller coaster blood glucose spikes in people who have trouble regulating blood sugar (and there are are millions in the US alone who don't even know they have that problem). That is one of the biggest lies perpetuated on overweight and/or diabetic patients, which leads to further weight gain and/or diabetic complications, because they can't keep their BG/insulin in a normal range on the prescribed high carb diets.

Those GI/GL tests were done on people with healthy BG/insulin-sensitive systems to begin with, so are not valid for those with impaired or over- burdened BG regulatory systems.

I have tested my BG with my glucose meter (calibrated against my labs test results) after eating many, many foods, and fiber or not, starch converts to glucose quite fast with or without fiber. I find little difference in my BG readings between whole grain and white flour, it's always much too high and starts a roller coaster effect. Starchy veggies are much the same.

The highest BG reading I have ever had from a food came from a measured serving of Cheerios cereal, supposedly a "whole grain" and emblazened with the AHA seal of approval. I now view whole grain cereals as predigested starch, a veritable bowl of sugar.

Whole grain is a misnomer anyway. What is called whole grain is more often actually "whole meal" or "whole grain flour". If the grain is crushed, milled, flattened, rolled, cut, etc., it isn't whole anymore.

Stephan Guyenet said...

Anna, the more time goes on, the less I'm convinced of the glycemic index/load thing.

My Year Without said...

Great information, here. I agree with your comment, "In my opinion, the trail of metabolic destruction that has followed Westerners throughout the world is probably due in large part to wheat and refined sugar." I believe that to be true because those ingredients are in just about everything--all those packaged products. I have gone without refined sugar all year, so far, and it has led to going without refined flour, as well. It is unbelievable to see all the refined ingredients in packaged products. There has to be a link, like you said, between failing health and refined products, especially the AMOUNT of refined products in packaged foods. I have found personally that it is more nutritionally rewarding to eat whole foods, but it is definitely not as convenient!

Kayaman said...

If the GI/GL isn't that important, than what's so special about sugar that causes hyperinsulinemia? Is it the fructose component, or the large quantity easily consumed in one meal? Something else, mabye?

Stephan Guyenet said...

My Year Without,

It really is ubiquitous, isn't it? I think that's why it's so valuable to have data from cultures that don't eat it.


I'm not completely abandoning the GI/GL just yet. I'm still open-minded to the possibility that it could play a role. I just don't think the totality of the data are convincing at this point.

Sugar in the large amounts we consume in industrial nations has a large load of quickly digesting glucose and fructose. That could be the key. Or it could be something neurological, for example the sweet stimulus is so large it tricks our appetite centers. I'm not really sure what the mechanism is, I've simply noticed that where sugar goes, disease follows.

Although to be honest, it usually shows up at the same time as wheat and other processed foods. I don't think sugar alone is sufficient to cause the diseases of civilization. The Kuna eat plenty of refined sugar and they're healthy. But they don't eat wheat and not much grains in general.

If I had to guess, I'd say it's a contributing factor but not sufficient to cause disease by itself.

Dave said...

Hi Stephan. Here's some wild-assed speculation...

I agree that the "carbohydrates drive insulin drives fat storage" is an oversimplification. The details of the process are interesting. LPL activity is considerably delayed from the initial insulin signal. LPL activity in adipose tissue peaks around 4 hours or so from the initial insulin spike. This corresponds nicely with the the peak of plasma chylomicron density from dietary fat; and though I don't know off the top of my head, I'll bet VLDL from the liver carrying triglycerides manufactured from dietary carbohydrate also enter in around the same time window. By this time, blood glucose has returned to nearly baseline. Peter has some excellent speculation on the Hyperlipid blog as to why evolution designed us this way, essentially to separate lipoproteins from elevated blood glucose, since the glucose damages the lipoproteins. It is these damaged lipoproteins which are implicated in the development of CVD due to the inflammatory reaction they cause.

So that in itself is a nice story, and is consistent with the absence of CVD in Kitavans. Their diet is not only relatively low fat, but a lot of the fat (presumably from coconuts) is MCT, hence does not require chylomicron transport. All in all, one would presume that the Kitavans would have lower numbers of lipoprotein molecules than would Westerners (who eat large relatively larger amounts of fat with their carbs). Less lipoprotein, less potential for damage. Wild speculation: Kitavan sources of carbohydrates are also high in anti-oxidant (and anti-glycation?) micronutrients, potentially extra protection. All of these aspects may lend further protection against the effects of smoking (not the least of which, I imagine, is increased oxidative damage to lipoproteins).

Plasma non-esterifed fatty acids (NEFA) cause insulin resistance, particularly in muscle (the main "sink" for dietary glucose). These would also be peaking around the time that chylomicrons/VLDL/LPL peak. If you eat fat with your carbs, you may be more insulin resistant for the next meal. Kitavans eat relatively low-fat, and a lot of the fat they do eat (MCTs) gets a ride straight to the liver, and much of that is presumably used for energy (the liver likes fat). So less NEFA, less insulin resistance, better glucose tolerance.

Another aspect: Vitamin C and glucose compete for binding sites, being chemically very similar. Vitamin C, amongst other things, strengthens connective tissue (like the inside of arteries). High intake of Vitamin C would be protective against some of the damages from dietary carbohydrate. I'll bet the Kitavans get a LOT of Vitamin C.

Omega-6/Omega-3 ratio is too coarse of a measure. Plant-based omega-3 (ALA) is inefficiently converted to its biologically useful form (in humans) DHA. The rest is just fodder for increasing oxidative stress, including oxidative damage to lipoproteins. Do the Kitavans get a fair amount of DHA from seafood (the requirement is actually quite small, I believe)?

We discussed a bit on my blog the possibility that grains could contain anti-nutrients that interfere with insulin signaling. I also mentioned in my last comment that some wild tubers (yams and sweet potatoes?) may have the opposite effect, actually helping with insulin control. I don't know the mechanism, but increasing insulin sensitivity and correspondingly glucose clearance sounds reasonable.

Let's not overlook the effects of psychosocial stress. Most primitive societies live rather low-stress lives (sort of ironic that we call them "primitive", isn't it). We all know chronic stress is bad. Increasing insulin tends to temporarily depress stress hormones; insulin also lights up the same brain reward centers which are studied in the context of addictive disorders. Maybe it's no surprise that "comfort foods" are almost invariably high-carb. A low-stress lifestyle may mitigate overconsumption of carbohydrates, allowing normal energy regulation mechanisms to work properly. Additionally, normal levels of stress hormones probably make for better glucose tolerance, lower insulin levels, etc.

Stephan Guyenet said...

Hi Dave,

Thanks for your thoughts. I've also been following Peter's recent posts with great interest. It definitely makes a case for saturated fat, or more precisely, avoiding polyunsaturated fat if you eat carbs.

One thing that bugs me about the argument that one must eat low fat on a high-carb diet to avoid insulin resistance is that there are empirical counter-examples. The Kuna of Panama don't seem to get the diseases of civilization, including overweight, even though their diet composition is similar to an American's (their fat is more saturated however). They get most of their carbs from plantains and cassava, but some from fruit, corn and sugar as well. Not much grain intake, virtually no wheat. I just don't think the answer is to be found in macronutrients. Why would our bodies have evolved to tolerate certain ratios of macronutrients and not others? The best design would accomodate whatever combination of food sources came along, and they did come along in every combination.

I can understand the argument for limiting polyunsaturated fats in a high-carb diet. I'm not aware of any non-industrial populations that are healthy on a high-carb, high-polyunsaturate diet. That's not to say there are none, but I haven't come across them.

I don't know what the Kitavans' lipoprotein numbers are, but their total cholesterol is about 30% lower than a Swede's on average. HDL is a bit lower, LDL is lower and trigs are a bit higher than Swedes. I published some of the numbers a few posts ago.

Kitavans probably do get a lot of vitamin C from all the fruit they eat. They also have a relatively high DHA intake, from the 4g of fish fat they eat per day.

I haven't heard about the effects of sweet potatoes on insulin sensitivity, but I don't think you have to invoke that to explain Kitavans' carb tolerance. For example, in the Papua NG mainland they eat plenty of sago palm, which is pure starch. The Kuna get their starch from plantains. They are both lean and healthy. It seems like the outlier is wheat rather than sweet potatoes.

About the stress factor, I believe Lindeberg made a brief attempt to address that. He says that on Kitava "Suppressed aggressiveness seems common. Intense fear of sorcery clasps everyone." I think it's very hard to determine a culture's stress level from a few months among them, so it's still possible that they have less stress. But at the very least, the answer is not straightforward.

Kayaman said...

Combining a couple ideas from above:
1) Chylomicrons+VLDL peak a few hours after a meal
2) Americans have a propensity to snack (on carbs) a few hours after a meal

I can certainly see how this would be a very bad situation indeed!

Anonymous said...


I'd love for you to expound on your skepticism on GI/GL. One thing that popped out at me from Wikipedia:

# GI takes into account only the resulting glucose in blood chemistry, with no consideration of the effect of other sugars, particularly fructose. An example is Coca-Cola. While already having a medium GI rating, Coca-Cola actually contains even more sugar than is otherwise being accounted for by the GI, as it contains HFCS 55, meaning that for each serving, the 55% fructose is not being accounted for. High levels of fructose cause the same blood sugar spike as a similar amount of glucose[citation needed]. While fructose does not produce the large insulin response of glucose [2], it can cause its own set of health problems, due to an increase in plasma lipids. [3]

If this is true, isn't GI largely useless?

Dave said...

Hi Stephan. I'm going to have to work up a blog post on this and related topics. Thinking about boundary cases like the Inuit and Kitavans is interesting and informative.

I started poking around to try and find other cases, e.g. is there a population that eats a lot of wheat and little or no sugar, or a population that suffers from "diseases of civilization" without wheat. One interesting instance is the Aztecs, which you can read about here:

The relevant stuff is down in the middle of the article. Here's some interesting excerpts:

"The Badianus Manuscript index provides commentary on approximately 100 medical conditions. Of these, 10% are nutrition or foodintake related and include reference to conditions of angina, constipation, dental problems (tartar removal), dysentery, dyspepsia-indigestion, fatigue, gout, heart (overheated), hemorrhoids and lactation difficulties. No references appear in this manuscript that could be equated with beriberi, pellagra, rickets or scurvy, or to medical-nutrition related conditions such as cancer, diabetes or stroke...

Other texts document nutritional problems among the Aztecs, including heart disease, kwashiorkor(5) and obesity. Ortez de Montellano (1990) wrote that traditional Aztec diet apparently was low in saturated dietary fats, given only modest intake of meat and the lack of dairy products before the Spanish conquest. He suggested that such a pattern, coupled with a presumed high level of physical activity among Aztec workers, would minimize cardiovascular disease. Yet his review of surviving Aztec documents identified heart ailments such as angina and "overheated heart."...

The Aztecs had a sophisticated vocabulary for obesity and locations of specific fat deposits: cotztzotzol--soft, drooping fat and skin deposited along the calf; eltzotzolli--flabby, fatty tissue and skin deposited across the chest; ititzotzolli--soft, bland fat and skin in the hypogastrium region; puchquiyot| very flabby fat widely distributed over the body; and quechtzotzol--a flabby, double chin (Lopez Austin, 1988, p. 259)."

I haven't had a chance to think about this in detail, and would like to uncover other similar cases as well. But the similarities and differences (e.g. heart disease and obesity but no cancer or diabetes) are thought-provoking. Corn, of course, comes with its own complement of anti-nutrients, e.g. lectins. Comparison with wheat for similarities should be educational.

Chris said...


this is interesting and possibly relevant to the GI / GL debate

I've just got Aragon's book - Girth Control - which looks quite interesting

Stephan Guyenet said...


That is an excellent article, thanks for sending it my way. I'll have to check that book out.

Stephan Guyenet said...

Hi Dave,

Very interesting. I wonder how common those ailments were?

I've read a bit about S American archaeology, and I've picked up a few interesting facts. Corn was not a widespread dietary staple until about 800 AD. When it became a staple, skeletons showed the characteristic signs of grain adoption: shortened stature, osteopenia, osteoporosis, narrower pelvic inlet, tooth decay, skeletal signs of anemia. Probably due to the increase in dietary anti-nutrients and the replacement of nutrient-dense foods.

I do believe it's possible to be overweight without eating wheat, by the way. It just seems like rampant, society-level obesity only occurs in wheat-based cultures.

I agree that boundary cases are fascinating, that's exactly why I've latched on to the Kitava study.

Anonymous said...
This comment has been removed by a blog administrator.
jwgn said...

Check out the Nuba tribe in the Upper Nile area. They are truly models of physical perfection. Diet staple is sorghum cereal--fermented.

Also sesame. Of course, this isn't exahaustive.

TCO348 said...

Maybe total calories drives insulin sensitivity?

Unknown said...

You said: "Western food habits." and later specified "wheat and refined sugar." But I'm confused: You're saying it's not GI, it's just wheat and refined sugar: But what else do these things have?

The answers that spring to mind are:
anutrients (that are not effectivly processed in industrial foods?)
lack of nutrients
The only other thing that might factor is, as someone menetioned, snacking, as there are some studies linking Intermittent Fasting to better insulin management if I'm not mistaken.
Do you know anything abou the Kitava habits in terms of meal frequency and size?

Do you think these are the factors that you're getting at when you say "Western food habits" and "sugar and wheat"?

BTW, your blog is a jewel, thanks so much -I can't wait to comb through the rest of it. (I do realize the original post is 2 years + but... oh well, I hope you can reply.)

Unknown said...

Sorry, this should have been in the last post: What kind of studies would you like to see to prove an alternative?
Have you seen this?

Unknown said...

The wheat piece is probably the big clue - but perhaps not the way you think. The big shift in western diets occurred with factory farming from the 1950's. This led to a massive shift in Omega3/6/9 ratios in the diet. A diet rich in O3's reduces insulin sensitivity and leads to better functioning mitochondria. Factory farming shifted to grain based foodstocks for most animals - which massively increased the O6's at the expense of the O3's. One classic example is chicken. < 1950's a chicken breast contained over 120mg of O3's, nowadays < 20mg. Grassfed all the way for dairy and meat.

Unknown said...

The controversies over cardio for fat loss are endless: steady state versus intervals, fed versus fasted, long and easy versus short and intense, and so on. Obviously there is a lot of interest in cardio training and how to do it right. Sadly, most people are still doing 2 things terribly wrong and it’s killing their results…… As best as I can figure, there are two major reasons why people are still mucking up their cardio programs for fat loss.


Most people aren’t burning enough darn calories.

Why? Well, I guess they are too busy worrying about the “proper” type of exercise (which machine or activity), the mode (steady state or intervals), the “optimal” ratio of intervals, or the “best” duration.

Some people coast along on the treadmill at 2.3 miles per hour or some similar sloth-like pace and they think that just by hitting a TIME goal, such as 45 or 60 minutes, that with “X” duration completed, they are assured to get the results they want.

On the other extreme, we have folks who have found or created some mega-intense, super-duper short training protocol like the “4-minute wonder workout from Japan.” Just because the workout is high in intensity and it is performed in intervals, they too think they are assured to get the results they want.

What’s missing in both cases is the realization that total fat loss over time is a function of total calories burned over time (assuming you don’t blow your diet, of course).


2 Cardio Mistakes You're Still Making